Closed-form Filtering for Non-linear Systems

arXiv:2402.09796v1 Announce Type: cross
Abstract: Sequential Bayesian Filtering aims to estimate the current state distribution of a Hidden Markov Model, given the past observations. The problem is well-known to be intractable for most application domains, except in notable cases such as the tabular setting or for linear dynamical systems with gaussian noise. In this work, we propose a new class of filters based on Gaussian PSD Models, which offer several advantages in terms of density approximation and computational efficiency. We …

abstract application arxiv bayesian cases cs.lg cs.ro current distribution domains filtering form hidden linear markov noise non-linear state stat.ml systems tabular type work